MODEL ADOPTION #4031: REJECT — LEDGER BENCH: Apply filter: Min confirmations >= 5 — improves win rate from 32.5% to 34.0% — E8 Intelligence Research
Year: 2026
Type: preprint
Abstract: DECISION: REJECT ELEMENT: No element survives — the +1.5pp win-rate gain is not a concrete, testable edge compatible with E8 geometry timing; it is a statistical artifact of trade-count reduction, not a timing improvement. EXPECTED IMPACT: Win-rate may rise ~1.5pp, but total PnL likely flat or negative because the filter removes early entries that carry higher R-multiples in fast regimes (as Merlin noted). The backtest n=1816 is regime-mixed, and the live window shows no confirmation-dependent loss pattern — the filter is not targeting a timing flaw. RISK: Over-fits to a low-confirmation-loss cluster in the backtest; in live trending markets, waiting for 5 confirmations delays entries, causing missed breakouts and slippage on late fills. Also reduces sample size, increasing variance and false confidence. TEST: Before any adoption, run a 90-day shadow replay on all 7 instruments, logging every signal with confirmation count. Must pass: (1) filtered vs unfiltered average R-multip Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
Author: Caldin, Andrew Stewart
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7221661300
Cites:
Cited by:
Related to: